Mistral AI

Mistral AI is a French AI startup founded by former researchers from Google’s DeepMind and Meta Platforms. The company aims to make AI useful for enterprises by leveraging publicly available data and contributions from customers. Mistral AI recently released its first large language model called Mistral 7B. This model is available for free and can be downloaded via various means, including a 13.4-gigabyte torrent. 

It outperforms other models of similar size and is released under a highly permissive license that allows unrestricted use and reproduction, as long as proper attribution is given. Mistral AI’s ambition is to become a leading supporter of the open generative AI community and bring open models to state-of-the-art performance.

Key Features:

Performance: Mistral 7B outperforms other models of similar size according to standard benchmarks.

Cost-effective: It offers similar capabilities as larger models like Llama 2 but at a considerably smaller compute cost.

Faster inference: Mistral AI achieves faster inference using grouped-query attention (GQA).

Longer sequences: It handles longer sequences at a smaller cost using Sliding Window Attention (SWA).

Mistral AI’s model has gained attention due to its open nature and high performance. It is released under a highly permissive license, allowing it to be used by hobbyists, corporations, or even government entities1. The company aims to become a leading supporter of the open generative AI community and bring open models to state-of-the-art performance.

Please note that while the model is free to use, Mistral AI also offers commercial solutions with additional features and support.